Joint Bayesian Calibration of Frontal Ablation and Surface Mass Balance in Global Glacier Models

Modeling frontal ablation of marine-terminating glaciers is critical for accurately modeling mass change but remains challenging. We present a hybrid, physically consistent and numerically efficient modelling framework that integrates the Simple Estimator of Retreat Magnitude and ice flux (SERMeQ) into two coupled global glacier models (OGGM and PyGEM). The adaptive particle‑batch smoother, a Bayesian data assimilation method, is used for joint calibration of all parameters. The coupled framework simulates monthly length changes along flowlines and mass‑balance components of individual glaciers. Applied to 71 marine-terminating glaciers across Svalbard during the period of 2000–2019, the model simulates distinct seasonal variations and annual length changes of glacier calving fronts, while reproducing decadal climatic mass balance estimates well. However, frontal ablation tends to be overestimated possibly due to uncertainties in initial geometry and modelled ice velocities. Overall, the good alignment of the regional distributions highlights that the modeling framework, including the efficient joint Bayesian calibration, provides a promising tool for large-scale frontal ablation modeling in glacier evolution models.